Lidar observations during RV <italic>Polarstern</italic> cruises across the Atlantic
The temporal development of the range-corrected signal (i.e. the uncalibrated
attenuated backscatter signal) of the autumn transit cruise PS95 is shown in
Fig. b. RV Polarstern departed on 29 October 2015
from Bremerhaven (Germany) and arrived on 1 December 2015 at Cape Town
(Republic of South Africa). The first days of this cruise were characterised
by low-level clouds and rain indicated by high signals (white colours). On
9 November (33∘ N), the lidar could detect a lofted plume of Saharan
dust above the MBL between 600 m and 3 km height. From 12 November
(24∘ N), increasing depolarisation in the MBL could be observed
(Fig. c) resulting from deposition and down-mixing of dust
from higher altitudes. The dust top height decreased from 2.8 km on
11 November down to 1.5 km on 13 November. About noon on 14 November 2015, a
new dust plume with a lower volume depolarisation ratio and a dust top height
of 3.5 km was observed. RV Polarstern steadily moved towards the
equator so that the dust region was left behind in the night from 17 to
18 November (3∘ N). After entering the Southern Hemisphere on
19 November, marine stratocumulus clouds occurred frequently. Around noon on
23 November (10∘ S), minor traces of dust between 1 and 4 km could
be observed again. Considering HYSPLIT trajectories (not shown), these
depolarising layers could consist of dust from the Kalahari Desert. From
24 November (12∘ S) onwards, the sky was mostly overcast. At the end
of the cruise, on 29 and 30 November, almost pure marine conditions could be
observed. The 500 nm aerosol optical thickness (AOT; Fig 2a), measured
with a Microtops sun photometer, ranged around 0.1 on the Northern
Hemisphere, increased in the dust-influenced northern tropics to around 0.5,
and decreased below 0.1 in the Southern Hemisphere.
Same as Fig. but for the spring cruise
PS98 from Punta Arenas to Bremerhaven. Due to the failure of the
1064 nm channel, the 532 nm range-corrected signal is shown in the
middle panel. The white bar indicates the time period of the case study discussed in
Sect. .
![]()
The spring transit cruise PS98 started on 11 April 2016 in Punta
Arenas (Chile) and ended on 11 May 2016 in Bremerhaven. Time series
of the range-corrected signal and volume depolarisation ratio at
532 nm are shown in Fig. . Because of a failure of
the 1064 nm photomultiplier tube (PMT), no measurements at this
wavelength were available. After starting regular measurements in
the night from 12 to 13 April, the weather was dominated by
clouds. In the night from 14 to 15 April (40∘ S), thin
depolarising layers at around 2.5 km could be observed. According
to HYSPLIT backward trajectories (not shown), the air mass
originated from the Patagonian region; thus, the layers could
contain traces of Patagonian dust. Observations on 16 and 17 April
were dominated by low clouds and rain. From 22 April
(12∘ S) to 25 April (4∘ S), a lofted dust plume
between 1.5 km and 3.5 km height could be detected. Crossing the
Intertropical Convergence Zone (at around 5∘ N) on the 27
and 28 April, thunderstorms, rain showers, and clouds with low base
heights were predominant. After leaving this region, the lidar
observed Saharan dust above the marine boundary layer again. The
bottom height of the dust layer decreased from 1.5 km on 22 April
down to around 600 m on 30 April. In the afternoon of 1 May
(23∘ N), RV Polarstern left the dust region.
After a short stop at the port of Las Palmas (Gran Canaria, Spain)
on 3 May, the cruise was continued towards the European continent
and the aerosol conditions were more and more influenced by
anthropogenic sources. From 6 May onwards, mostly overcast sky with
small cloud gaps was predominant. The AOT at
500 nm (Fig. 3a) showed a similar
zonal behaviour as on the PS95 cruise. The AOT at 500 nm was below
0.1 in the Southern Hemisphere, except for 17 April, and steadily
increased to the maximum of 0.37 on 30 April. After leaving the
dust-influenced region, the AOT ranged between 0.1 and 0.2 in the
Northern Hemisphere.
Regular cruises across the Atlantic Ocean from north to south in the
northern hemispheric autumn and from south to north in the northern
hemispheric spring provided a large amount of lidar data over the
Atlantic. Dust has been regularly observed in the northern tropics and subtropics west of
the Saharan desert. The AOT at 500 nm, measured with a Microtops
sun photometer, has been slightly higher in
the Northern Hemisphere than in the Southern Hemisphere, which
indicates a higher aerosol load in the former.
Marine conditions during PS95: time series of the Microtops
sun-photometer-derived AOT at 500 and 440/870 nm Ångström
exponent (a), 1064 nm range-corrected signal (b), and
532 nm volume depolarisation ratio (c). Vertical white lines
indicate the signal-averaging period for profiles shown in
Fig. . The black star in the cruise map shows the
location of RV Polarstern during this period.
![]()
Profiles averaged for 30 November 2015, 01:15–02:30 UTC.
Backscatter coefficient and depolarisation
ratios are smoothed with
127.5 m vertical length. Extinction coefficients, lidar ratios, and
Ångström exponents are smoothed with 127.5 up to 242 m and
afterwards with 367.5 m. Meteorological data from GDAS1
(30 November 2015, 00:00 UTC) and radio sounding measurements
(29 November 2015, 12:00 UTC) are also presented.
![]()
Case studies
Three night measurements from PS95 and PS98 were selected to present
typical atmospheric conditions by means of a detailed discussion of
the optical properties in the MBL and in lofted layers. First,
almost pure marine conditions with an overlying dried marine aerosol
layer during the autumn cruise PS95 are discussed. Second, a case
study on the same cruise but with Saharan dust near the Canary
Islands is presented. Third, a case during the spring cruise in 2016
(PS98) with Saharan dust and biomass-burning aerosol mixtures near
the Cabo Verde islands is shown. These three case studies are
marked with black stars on the cruise tracks
(Fig. ) and with white
lines in the cruise overviews (Fig. , ).
PS95 – marine aerosol conditions
On 29 and 30 November 2015 at the end of the cruise PS95, clean conditions
could be observed near Cape Town. In this area, the dominant aerosol was of
marine origin according to CALIPSO aerosol classification .
In Fig. , the time series of the range-corrected
signal at 1064 nm and the volume depolarisation ratio at 532 nm from 29 and
30 November 2015 are shown. Additionally, the AOT at 500 nm and the
Ångström exponent at 440/870 nm by sun-photometer
measurements are shown in the upper panel. Mean AOT at 500 nm of
0.09 ± 0.01 on 29 November (0.09 ± 0.02 on 30 November) and an
Ångström exponent of 0.08 ± 0.02 (0.23 ± 0.08) clearly
indicate marine conditions for remote oceanic areas, not influenced by
continental aerosol sources. In these regions, the AOT at 500 nm is
typically below 0.1 and the Ångström exponent less than 0.4
.
The time series of the volume depolarisation
ratio (Fig. 4b) shows a thin layer
of enhanced depolarisation at the top of the MBL at 300–400 m.
This layer consists of dried marine particles and will be discussed
later in this section.
Mean profiles of the measured optical properties are shown in
Fig. for 30 November 01:15–02:30 UTC. In the right
panel, GDAS1 and radio sounding profiles are shown. The temperature inversion
and decrease of the relative humidity (RH) as well as the strong decrease of the
backscatter signal suggest the MBL top height at about
300 m. Within the MBL, the lidar ratio was 23 ± 2 sr at 355 and
23 ± 1 sr at 532 nm, which agrees with results during the second
Aerosol Characterization Experiment ACE-2 S532
23 ± 3 sr; and are slightly higher than results of the
Saharan Mineral Dust Experiment SAMUM-2a S532 18 ± 4 sr
and S355 18 ± 2 sr;.
NOAA HYSPLIT backward trajectories for 4 days ending at the
position of RV Polarstern (31.48∘ S,
14.51∘ E; marked by the black star) on 30 November 2015,
02:00 UTC, at 300, 600, and 1000 m a.g.l.
![]()
(a) Column-integrated dust concentration (g m2) and
3000 m wind on 11 November 2015, 12:00 UTC, from the BSC-DREAM8b model
(Dust Regional Atmospheric Model), operated by the Barcelona Supercomputing
Center (http://www.bsc.es/ess/bsc-dust-daily-forecast, last access: 14
November 2016). (b) The 7-day NOAA HYSPLIT backward trajectories ending
at the position of RV Polarstern on 11 November 2015, 20:00 UTC
(24.27∘ N, 19.23∘ W). The position of RV
Polarstern is marked by the black star.
![]()
The special highlight in this case study is the increase of the
depolarisation ratio at the top of the MBL, whereas the lidar ratio
within this layer is low, 16 ± 1 sr (355 nm) and
13 ± 3 sr (532 nm). Thus, this layer cannot consist of
biomass-burning aerosol or dust mixtures. The particle
depolarisation ratios at 355 and 532 nm are around zero in the MBL
(large, spherical particles) and increase from 300 to about
450 m, shortly above the MBL top. After this peak, the
depolarisation decreases to about zero again. Considering the
profiles of relative humidity and temperature, a correlation with
the relative humidity is obvious. The relative humidity decreases
from about 90 % near the ground to under 20 % above 600 m.
In the layer the RH is about 50 % according sounding data and
about 40 % according to GDAS1. Simultaneously the temperature
increases. HYSPLIT backward trajectories
(Fig. ) indicate that the air
parcels arriving at 300, 600, and 1000 m had only been carried over
the South Atlantic Ocean the last 7 days; thus, it can be assumed
that the air mass contains mostly marine aerosol, e.g. sea salt. Sea
salt aerosol exists as dry particles at low relative humidity. Since
sea salt is hygroscopic, the salt particles absorb water to form
droplets when the RH exceeds the deliquescence relative humidity,
which is around 70–74 % depending on the composition of the sea
salt . If the RH decreases to the crystallisation
relative humidity 45–48 %;, the particles
crystallise from the droplet. At a RH above the crystallisation
relative humidity, the sea salt particles are in solution with
water and show low values of δ≈3 %
. When the RH is below the crystallisation
relative humidity, the sea salt particles crystallise and exist as
non-spherical particles due to the cubic shape of NaCl, the main
constituent of sea salt aerosol . As
non-spherical particles they cause higher depolarisation ratios. In
this case, dried sea-salt particles caused depolarisation ratios up
to 9 % at 532 nm and 10 % at 355 nm. Previous studies
showed similar results. measured high
depolarisation ratios (≈10 %) at 532 nm in the lower
atmosphere associated with sea breeze events in the coastal area of
Tokyo Bay. During the Saharan Aerosol Long-range Transport and
Aerosol-Cloud-Interaction Experiment (SALTRACE) winter campaign 2014 at Barbados,
detected an increase of the particle
depolarisation ratio up to 12 % at 355 nm, 15 % at 532 nm,
and 10 % at 1064 nm when the RH drops below 50 %.
observed low depolarisation ratios (< 5 %)
at 532 nm over a wide range of relative humidities, whereas
δpar>10 % was measured at low RH (< 50 %) in
air masses which had passed over the Pacific Ocean. In a laboratory
chamber experiment, found linear depolarisation
ratios at 532 nm of 1 ± 0.1 % for droplets,
8 ± 1 % for sea salt crystals, and 21 ± 2 % for
NaCl crystals.
Thus, we can conclude that marine particles were transported above the MBL
top, dried, and crystallised and therefore cause a high particle
depolarisation ratio even though the backscattering is low compared to the
MBL.
This case confirms that marine aerosol can cause depolarisation in the lidar
signal when RH is low. Without considering this property of marine aerosol,
aerosol layers above the MBL causing depolarisation may be falsely
classified. Automatic classification algorithms like the ones for CALIPSO,
EarthCARE, and other lidars should take these feature into account, if
relative humidity measurements are available, to not misclassify these
aerosols as, for example, mixed dust.
First dust event observed during PS95: time series of the Microtops
sun-photometer-derived AOT (500 nm) and 440/870 nm Ångström
exponent (a), 1064 nm range-corrected signal (b), and
532 nm volume depolarisation ratio (c). Vertical white lines
indicate the signal-averaging period for profiles shown in
Fig. . The black star in the cruise map shows the
location of RV Polarstern during this period.
![]()
PS95 – Saharan dust
When RV Polarstern approached the Canary Islands during the autumn
cruise 2015, the first dust plume was observed in the evening of 10 November
at around 28∘ N. The dust could be measured until 14 November.
Figure a presents the column-integrated
concentration on 11 November 2015, 12:00 UTC, from the BSC-DREAM8b
model, operated by the Barcelona Supercomputing Center.
The increased column dust load above the Atlantic at the position of
RV Polarstern is illustrated by dark green colour.
According to HYSPLIT backward trajectories
(Fig. b), the air mass measured on
11 November 20:00 UTC originated from the Saharan desert. Only air
masses that arrived at 3 km had been carried also over European
areas in the last 7 days.
The range-corrected signal at 1064 nm and the 532 nm volume depolarisation
ratio of the first dust plume are shown in Fig. .
Additionally, sun-photometer measurements from 11 and 12 November are given.
The sun-photometer-derived Ångström exponent at 440/870 nm is 0.13
on 11 November and 0.08 for the day after. The daily averaged AOT at 500 nm
for these days is 0.38 (11 November) and 0.58 (12 November). The dust layer
reached heights about 3 to 3.5 km on 11 November and slightly descended
towards 14 November. From 13 November lofted layers between 2 and 3.5 km
with a lower depolarisation ratio than the first dust plume could be
observed.
Profiles averaged for 11 November 2015, 19:30–21:00 UTC. Dust
fraction calculated following . Radio sounding profiles from
11 November 2015, 12:00 UTC, and GDAS1 profiles (11 November 2015,
18:00 UTC) are presented in the right panel.
![]()
Averaged profiles of the measured optical properties and radio
sounding and GDAS1 profiles of temperature and relative humidity are
shown in Fig. for 11 November
19:30–21:00 UTC (white frame in Fig. ).
Backscatter profiles show an increased backscatter coefficient at
all wavelengths from the MBL top (around 400 m) up to 2.8 km. The
backscatter coefficient at 532 nm is larger than at 355 nm,
whereas the extinction coefficient is wavelength independent. Even
though this is an atypical spectral behaviour, comparable
observations of higher 532 nm than 355 nm backscatter coefficient
have already been observed in dust layers near the Cabo Verde
islands , in the eastern Mediterranean at
Crete , and during the SHADOW (Study of SaHAran
Dust Over West Africa) campaign in Senegal .
According to , this spectral behaviour may be
caused by specific refractive index characteristics induced by the
chemical composition of the particles. The mean lidar ratio at
532 nm (355 nm) in the height of the lofted aerosol layer is
53 ± 2 sr (61 ± 4 sr). The lidar ratio at 355 nm is
higher than at 532 nm, which results from the higher backscatter
coefficient at 532 nm and agrees with values found for dust during
the SHADOW campaign . Consequently, the mean
backscatter-related 355/532 nm Ångström exponent is negative
(-0.4 ± 0.1). Negative backscatter-related Ångström
exponents are generally found when scattering properties of dust are
modelled by assuming a spheroidal shape distribution. The values
then typically vary between -0.5 and -2 depending on the
assumptions of the spectral refractive index and the size and shape
distributions. The extinction-related Ångström exponent at
355/532 nm of 0.1 ± 0.5 and the backscatter-related
Ångström exponent at 532/1064 nm of 0.4 ± 0.1 are in
good agreement with values for dust measured during SAMUM-2b
Åext355/532≈ 0.22 ± 0.27,
Åbsc532/1064≈ 0.45 ± 0.16;.
Furthermore, the aerosol layer between 600 m and 2.8 km is
characterised by a nearly height-constant particle depolarisation
ratio of 29 ± 1 % at 532 nm and 25 ± 1 % at
355 nm. The increased particle depolarisation ratios indicate a
non-spherical particle shape and are in good agreement with values
found for pure dust during SAMUM-2a δ532par≈30 % and δ355par≈25 %;. The fraction of dust and
smoke can be estimated using a method described by
. Assuming a δ532par of
31 % for pure dust and 5 % for smoke, the fraction of dust
in this layer amounts to over 90 % (Fig. , panel 6) and can
therefore be considered as pure dust.
The MBL reached a height of about 400 m according to the backscatter
profile. Lidar ratios at 532 and 355 nm are 30 ± 3 and
30 ± 2 sr in the MBL, which are higher than the characteristic values
for marine aerosol (see marine case study Sect. ) and
suggest a mixture of marine aerosol with other particles. Particle
depolarisation ratios are also slightly higher, 9 % at 532 nm and
6 % at 355 nm, and indicate the mixing of dust into the MBL. Therefore,
also the MBL is influenced by the frequent dust emission in the Saharan
desert.
Complex aerosol layering with smoke and dust on PS98:
sun-photometer-derived AOT at 500 and 440/870 nm Ångström
exponent (a), 532 nm range-corrected signal (b), and
volume depolarisation ratio (c). Vertical white lines indicate the
signal-averaging period for profiles shown in Fig. .
The black star in the cruise map shows the location of RV
Polarstern during this period.
![]()
Averaged profiles for 29 April 2016, 20:15–21:00 UTC. Dust and
smoke fractions calculated following . Meteorological data
from GDAS1 (29 April 2016, 21:00 UTC) and radio sounding measurements
(29 April 2016, 15:00 UTC). Layers are marked grey.
![]()
PS98 – mixed aerosol layers
During the spring cruise PS98, extended aerosol layers with enhanced
depolarisation were observed near the Cabo Verde islands. The range-corrected
signal and volume depolarisation ratio at 532 nm as well as Microtops sun-photometer measurements on 29 April 2016 are shown in
Fig. . Sun-photometer measurements determined an average
AOT at 500 nm of 0.23 and an Ångström exponent of 0.9 for
440/870 nm.
An increased backscatter coefficient at both wavelengths indicates
aerosol layers between 0.9 and 3 km. These layers are separated
from the MBL, which reached a height of about 500 m according to
the increased backscatter signal and the GDAS1 and radio sounding
data. The mean lidar ratio at 355 nm is 22 ± 1 sr, the mean
backscatter-related 355/532 nm Ångström exponent is 0.9 ± 0.0,
and the mean particle depolarisation ratios are around zero at both
wavelengths in the MBL. These values are indicators of a pure marine
boundary layer without dust (see marine case,
Sect. ).
Mean profiles of the optical properties averaged from 29 April 2016
between 20:15 and 21:00 UTC are shown in
Fig. . Regarding the backscatter profile,
the aerosol-laden region above the MBL can be divided into five
layers. The layers extend from 0.9 to 1.2, 1.3 to 1.6, 1.7 to 2.2,
2.3 to 2.5, and from 2.6 to 3.0 km and are marked grey in
Fig. . The 532 nm near-range extinction
coefficient and lidar ratio was not reliable because of a
misalignment of the 532 nm near-range channel which does not affect
the Raman backscatter retrievals. The mean lidar ratio in the first
layer is 48 ± 4 and 46 ± 9 sr at 532 and 355 nm,
respectively. The mean backscatter-related 355/532 nm
Ångström exponent is 0.4 ± 0.1 and ranges between the
typical values of dust (0.16 ± 0.45, SAMUM-2b) and smoke
0.90 ± 0.26, SAMUM-2a;. The same
applies for the mean particle depolarisation ratio, which is
20 ± 2 % at 532 nm and 15 ± 2 % at 355 nm.
These values are in good agreement with values for dust and smoke
mixtures measured during SAMUM-2a
δ355,532par≈ 16 %;
and represent a dust fraction of 63 %, applying the method
described by , which is shown in
Fig. (panel 6).
The second layer extends from 1.3 to 1.6 km. The mean lidar ratio is
57 ± 7 and 63 ± 8 sr at 532 and 355 nm, respectively. Mean
backscatter and extinction-related Å at 355/532 nm amount
0.1 ± 0.1 and 0.4 ± 0.2, respectively. The mean particle
depolarisation ratio of 24 ± 2 % at 532 nm and 19 ± 2 %
at 355 nm suggests a mixture of depolarising dust and non-depolarising smoke
with a dust fraction of 77 % following .
In the third layer, the mean lidar ratio is 40 ± 6 sr at 532 nm and
50 ± 6 sr at 355 nm. The mean backscatter and extinction-related
Å at 355/532 nm is 0.4 ± 0.1 and 1 ± 0.4, respectively. This
indicates the presence of small, absorbing particles. The particle
depolarisation ratio, 13 ± 1 % at 532 nm and 9 ± 1 % at
355 nm, is small compared to the other layers. Therefore, it can be
concluded that this layer contains a significant amount of non-depolarising
particles like smoke. According to the fraction of dust
is only around 35 %.
Between 2.3 and 2.5 km, in the fourth layer, the particle
depolarisation ratio rises again (18 ± 1 % at 532 nm,
14 ± 3 % at 355 nm). The mean lidar ratio is
40 ± 11 sr at 355 nm and 42 ± 3 sr at 532 nm. The
mean backscatter and extinction-related Å at 355/532 nm is
0.26 ± 0.06 and 0.08 ± 0.80. According to the fraction of dust in this layer is around
55 %.
The fifth layer is characterised by a high lidar ratio up to 88 sr at
532 nm and 68 sr at 355 nm and high backscatter and extinction-related
Ångström exponents of 0.4 ± 0.2 and 1.6 ± 0.6,
respectively. Particle depolarisation ratios decrease with increasing height
and amount around 16 % at 532 nm (13 % at 355 nm) at the lower edge
and 8 % at 532 nm (5 % at 355 nm) at the top. Using the dust–smoke
separation method described by , the dust fraction
decreases from 47 to 14 % within this layer.
Figure presents the HYSPLIT backward trajectories
for the last 10 days arriving at the position of RV Polarstern at
different altitudes. Air masses arriving in the MBL (500 m) had been carried
only over the Atlantic Ocean the last 10 days and therefore contained mostly
marine aerosol. Air masses arriving between 1 and 3 km were advected from
the African continent. The air masses arriving at 1 km height originated
from the Saharan desert and passed over active fire areas south-west of the
Saharan desert 6 days before arriving at the position of RV
Polarstern. Trajectories arriving at 1.5 and 2 km also passed over
the Saharan desert and active biomass-burning regions, but have never been
close to the ground. As investigated by , fires can
support the upward transport of dust into the free troposphere. A high amount
of dust in addition to biomass-burning aerosol could therefore also be detected in
these altitudes. In contrast, air masses arriving at 2.5 and 3 km were on
ground level over active fire regions for several days and could take up a
high amount of biomass-burning aerosol.
NOAA HYSPLIT backward trajectories ending at 29 April 2016
21:00 UTC at the position of RV Polarstern (15.32∘ N,
22.73∘ W; marked by the black star) at different altitudes.
Additionally, fires detected by MODIS on-board the Terra and Aqua satellites
are shown. Fires are accumulated over the 10-day period from 20 to
29 April 2016. Yellow colour indicates a large number of fires, and red dots
indicate a low number of fires in the considered period
(https://lance.modaps.eosdis.nasa.gov/firemaps, accessed: 24 February
2017).
![]()
Mean values of extinction coefficient, lidar ratio and particle
depolarisation ratio at 532 nm, and the backscatter-related Ångström
exponent at 355/532 nm (top down) for MBL (blue), elevated aerosol
layers (black), and dried marine layers (red) on PS95 (a) and
PS98 (b) from north to south. Error bars indicate the standard
deviation. MBL top height and extent of the elevated layers are shown in the
first row.
![]()
During this night measurement, five layers with different fractions of dust
and smoke could be detected. At the same time, the MBL was almost pure marine
without mixed-in dust or smoke particles. This case study shows that the
MBL is not always influenced by dust and smoke transport and different
aerosol types can occur at the same time above the Atlantic.
Statistical analysis
A statistical analysis of all Raman measurements with suitable weather
conditions and signal quality was performed to provide an overview of
latitudinal differences and characteristics of the different aerosol types
observed over the Atlantic. A total of 45 night measurements from PS95 and
PS98 were selected for analysis with respect to optical aerosol properties.
Each measurement was screened for separated aerosol layers. The MBL and, when
present, elevated aerosol layers and layers of dried marine aerosol, as
presented in the first case study, have been analysed separately. These
layers with enhanced depolarisation ratio directly above the MBL will be
named dried marine layers.
Time series
The MBL top height and the extent of analysed elevated aerosol layers and
dried marine layers are shown in the first row of
Fig. for PS95 (panel a) and PS98 (panel b),
illustrated with blue dots and black and red bars, respectively. The MBL top
height ranges between 300 and 900 m and shows no clear latitudinal trend.
Mean values of extinction coefficient, lidar ratio and particle
depolarisation ratio at 532 nm, and the backscatter-related Ångström
exponent at 355/532 nm are shown in the panels below. Blue dots
illustrate the MBL mean values derived from near-range measurements. Mean
values of the elevated aerosol layers are derived from far-range signals and
are illustrated with black dots, whereas mean values of the dried marine
layers are illustrated red. These mean values are derived from near-range
signals. An exception is the measurement at 22∘ S where the
far-range signal is used because of the height of the dried marine layer.
Error bars represent the standard deviation. Measurements at 355 nm are not
shown for the sake of clarity but show similar results.
Mean MBL lidar ratios during PS95 are around 25 ± 3 sr in the
northern latitudes and 20 ± 3 sr in the Southern Hemisphere.
In the region of dust, between 35∘ N and the equator, the
mean lidar ratio in the MBL is 30 ± 6 sr. The increased lidar
ratio is caused by down-mixing of dust from higher altitudes,
whereas the lidar ratio in the southern latitudes correlates with
pure marine values. Anthropogenic aerosol from the European
continent influences the MBL in northern latitudes; the lidar ratio
is therefore slightly higher than for a pure marine environment.
High particle depolarisation ratios up to 18 % in the MBL
confirm the presence of depolarising particles in the region west of
the Sahara, while it is below 1 % in the European-influenced
North Atlantic and the marine-dominated South Atlantic. During PS98,
no significant increase of the MBL lidar ratio between
35∘ N and the equator could be observed. The mixing of dust
into the MBL is therefore considered to be
negligible. This is confirmed by
a continuous low particle depolarisation ratio of less than 1 %
in the MBL throughout the whole cruise. This contrast to PS95 can be
explained by seasonal variations in the dust transport and
deposition processes over the Atlantic.
Elevated aerosol layers were mainly observed between 30∘ N
and 15∘ S and show a wide range of mean lidar and particle
depolarisation ratios, caused by different particle types in these
layers. The mean particle depolarisation ratio in elevated layers
decreases from around 30 % at 25∘ N towards the south
during PS95, which suggests an increasing mixing with other less
depolarising particles. At 1 and 12∘ S during PS95 and in
the upper aerosol layers at 15 and 19∘ N during PS98, the
lidar ratio is higher than the other days (64–88 sr) while the
particle depolarisation is low (< 10 %). This indicates a
mixture with other absorbing, non-depolarising particles like
biomass-burning aerosol. During PS98 at 5∘ S, 2500 km from
the African coast, the lidar ratio between 1.5 and 2.4 km is around 35 sr, which is considerably
lower than the characteristic values of pure dust
S532≈ 55 sr;. This may
result from marine aerosol transported upward by turbulent mixing
processes .
Latitudinal differences can also be seen in the course of the
backscatter-related Ångström exponent at 355/532 nm in the
MBL during PS95. The mean Åbsc355/532 is around 0
in the dust region, whereas it is around 1 in the northern and 0.5
in southern latitudes. This suggests a mixture of marine aerosol and
large dust particles in the MBL between 35∘ N and the
equator, whereas in northern and southern mid-latitudes the fraction
of smaller particles dominates. In elevated layers, mean
Åbsc355/532 ranges between -0.5 and 0.5. From 30
to 20∘ N, Åbsc355/532 is negative and in
the second part of the plume it becomes positive. According to
a low backscatter-related Ångström
exponent indicates an increased imaginary part of the refractive
index at 355 nm compared to 532 nm and therefore a higher
absorption at 355 nm than at 532 nm. This is a result from
different aerosol sources and particle properties. During the second
cruise, mean Åbsc355/532 ranges between -0.5 and
1.5. In dusty layers the mean Å is generally smaller than in the
MBL – an exception is the upper layer at 19∘ N. In this layer, a high amount of
small soot particles cause a high Ångström exponent.
Ångström exponents in the MBL do not show indications of
down-mixed dust during this cruise.
The most prominent feature of dried marine layers is the enhanced
particle depolarisation ratio of about 4–9 % compared to the MBL with depolarisation ratios below 3 %. Those values are similar to
previous observations by and
. Whereas the lidar ratio in the dried
marine layer measured during PS95 is slightly lower than in the MBL,
it is about 40 sr in the case measured during PS98 but shows high
uncertainty in the latter. Mean Ångström exponents are
around 0.5 during PS95 and around 1 during PS98, but do not show a
clear difference from MBL mean values.
Differences in optical aerosol properties between northern and southern
latitudes and the dust-influenced region west of the Saharan desert were
detected. Whereas the Northern Hemisphere is influenced by anthropogenic
pollution, southern latitudes are more likely to be influenced by marine
aerosols only. Nevertheless, pure marine conditions, not influenced by
aerosol originating from the continent, are rare and could only be observed
at the end of PS95 near South Africa. Mostly, low-level clouds at the top of the
MBL at the southern latitudes prohibited the lidar data analysis and thus the
evaluation of more cases of pure marine conditions. In about 65 % of the
cruise time in the Southern Hemisphere and 50 % in total during both
cruises, clouds along the cruise track did not allow lidar data analysis.
Lidar ratio as a function of the particle depolarisation ratio at
355 nm (a) and at 532 nm (b) from all analysed MBL
(blue), elevated aerosol layer (black), and dried marine layer (red)
measurements of PS95 and PS98. Coloured ellipses denote the different aerosol
categories.
![]()
Optical properties for particle typing
Mean values of optical properties of the MBL and elevated aerosol layers from
PS95 and PS98 (shown in Fig. ) are discussed to
illustrate the potential of aerosol classification using intensive optical
quantities. Similar to classifications shown by and
the lidar ratio at 355 nm (532 nm) is presented against
the particle depolarisation ratio at 355 nm (532 nm) for elevated layers
(black), MBL (blue), and dried marine layers (red). Error bars were omitted
for the sake of clarity. Coloured ellipses denote the different aerosol
categories.
A clear separation of marine and dust-influenced MBL measurements
can be seen. Pure marine MBL measurements show lidar ratios between
13 and 40 sr and particle depolarisation ratios less than 2.5 %
at 355 and 532 nm, whereas the particle depolarisation ratio of
dust-influenced MBL measurements ranges between 5 and 20 %, caused
by a significant amount of non-spherical particles in the MBL. The
lidar ratio within these layers also shows a tendency to higher
values with
increasing particle depolarisation, caused by dust particles.
Elevated aerosol layers can be divided into layers with a high
particle depolarisation ratio (20–30 %) and a lidar ratio of
about 50–60 sr, layers with a lidar ratio between 30 and 75 sr
and a moderate particle depolarisation ratio (< 20 %), and
layers with a high lidar ratio (> 80 sr) and a low particle
depolarisation ratio (< 10 %). The first category is considered
for pure dust cases, whereas the mixing with other non-depolarising
particles is the second category, named dusty mixtures. These
particles could be spherical marine particles or biomass-burning
aerosol. If the lidar ratio is higher than reference values of
pure dust (≈ 55 sr), the aerosol is considered to be soot;
a lower lidar ratio indicates a mixture with marine particles.
Layers in the third category with lidar ratios greater than 80 sr
and a particle depolarisation lower than 10 % are considered as
smoke dominated. Mean values at 355 and 532 nm show similar
results, although lidar ratios at 355 nm tend to be slightly higher
for all aerosol categories. As clearly seen, some of the mixed
aerosol states do overlap and a clear separation by the lidar ratio
and depolarisation ratio is not possible.
Therefore, to complete the picture of particle-type-dependent
optical properties, the backscatter and extinction-related
Ångström exponents at 355/532 nm and the
backscatter-related Ångström exponent at 532/1064 nm are
considered for particle type separation in addition to the lidar and
depolarisation ratio in Fig. . The backscatter
and extinction-related Ångström exponent as a function of
the particle depolarisation ratio (Fig. bottom
face) shows that the Ångström exponent is not a suitable
parameter for the separation of pure marine and dust-influenced MBL,
while a clear separation is possible considering the depolarisation
ratio (pure marine: δpar < 5 %; aerosol
mixtures in the MBL: 5 % < δpar < 20 %).
For elevated aerosol layers, a slight tendency towards negative
backscatter-related Ångström exponents at 355/532 nm with
increasing depolarisation ratio values can be seen
(Fig. a, c). The extinction-related
Ångström exponent at both depolarisation wavelengths is more
widely dispersed than the backscatter-related Ångström
exponent but shows similar patterns
(Fig. b, d). In the illustration of the
backscatter-related Ångström exponent at 532/1064 nm
against the particle depolarisation ratio at 532 nm
(Fig. e) no tendency to smaller Ångström
exponents with higher depolarisation ratio of the elevated aerosol
layers can be observed; thus, this parameter is obviously not
suitable for a distinction between the different aerosol types.
Considering the lidar ratio as a function of the backscatter and
extinction-related Ångström exponents
(Fig. right face), it again becomes obvious
that the Ångström exponent is a much less powerful parameter
for aerosol typing in a marine environment compared to the lidar
ratio and depolarisation ratio.
Resulting from the preceding investigations, we consider the lidar ratio
together with the particle depolarisation ratio as best indicators for
particle classification above the Atlantic. A clear characteristic in terms
of lidar ratio and Ångström exponent for the dried marine layers is
not visible. Further observations of those layers are needed to get a
comprehensive picture of dried marine aerosol properties.
The values presented above might be valuable information for new
aerosol typing schemes needing knowledge from marine areas at the
specific lidar wavelengths as, for example, for the upcoming
EarthCARE mission. The operated lidar will measure at 355 nm but
also requires information on the spectral behaviour of the optical
aerosol properties to obtain radiation closure, which is one goal of
this mission.
Three-dimensional illustration of the relation between lidar ratio,
depolarisation ratio, and backscatter and extinction-related Ångström
exponent. Blue dots represent MBL measurements, black dots elevated aerosol
layers, and red dots dried marine layers. Coloured ellipses denote the
different aerosol categories as in Fig. 14.
![]()